Export Reviews, Discussions, Author Feedback and Meta-Reviews

The "encoding of soft probabilistic constraints as hard constraints for symbolic search" may be more novel, but the brevity of sec 2.2 makes it hard to understand this in detail. Significance: This paper is tackling a hard and interesting problem, that of the unsupervised learning of programs; this generalizes unsupervised learning applied to vector data to other kinds of data (e.g.